数控机床热误差补偿模型稳健性比较分析

发布时间：2018-04-16 07:50

本文选题：数控机床 + 热误差　；参考：《机械工程学报》2015年07期

【摘要】：数学模型的精度特性和稳健性特性对数控机床热误差补偿技术在实际中的实施性影响不容忽视。对数控加工中心关键点的温度和主轴z向的热变形量采用多种算法建立了预测模型,对不同算法拟合精度进行分析。同时进行全年热误差跟踪试验,获得了机床在不同环境温度和不同主轴转速的试验条件下的敏感点温度和热误差值。以此为基础,对各种预测模型的预测精度进行比较验证不同模型的稳健性。结果表明,多元线性回归算法的最小一乘、最小二乘估计模型以及分布滞后模型在改变试验条件时预测精度下降,而基于支持向量回归机原理的热误差补偿模型仍能保持较好的预测精度,稳健性强。这为数控机床热误差补偿模型的选择提供了具有实用价值的参考,具有很好工程应用性。
[Abstract]:The accuracy and robustness of the mathematical model can not be ignored in the practical application of the thermal error compensation technology of NC machine tools.Several algorithms are used to predict the temperature of key points in NC machining center and the thermal deformation of spindle z direction. The fitting accuracy of different algorithms is analyzed.At the same time, the thermal error tracking tests were carried out throughout the year, and the sensitive point temperature and the thermal error of the machine tool under different ambient temperature and different spindle speed were obtained.On this basis, the prediction accuracy of various prediction models is compared to verify the robustness of different models.The results show that the prediction accuracy of the least square estimation model and the distributed lag model of the multivariate linear regression algorithm are decreased when the experimental conditions are changed.The thermal error compensation model based on the principle of support vector regression still has good prediction accuracy and strong robustness.This provides a practical reference for the selection of thermal error compensation model for NC machine tools and has good engineering application.
【作者单位】：合肥工业大学仪器科学与光电工程学院;
【基金】：国家自然科学基金资助项目(51175142)
【分类号】：TG659

【参考文献】